Vimana
The old Sanskrit epics tell of palaces that fly. Suppose an engineer took those stories the way they would take a client's brief, and priced them against the periodic table. The first thing they would notice is that the brief does not describe one machine. Air, vacuum and being carried each want a different answer. Here is that brief, costed mode by mode against what the shelf can supply.
The stories
In the Ramayana, one of the two great Sanskrit epics, the hero rides home aboard the Pushpaka vimana: a palace that moves through the sky at its owner's wish, roomy enough for a crowd, gentle enough that nobody aboard feels the ride. The Mahabharata, the other great epic, carries sky chariots of its own. How far back the material runs is argued over, and the arguments are worth reading on their own account. What is not in dispute is that generation after generation chose to carry the flying palace forward rather than drop it.
India is not alone. Persian tales carry flying carpets, Chinese stories tell of cloud chariots, and medieval Irish annals tell of ships seen sailing through the air. People in many places have looked up and told of riding the sky.
These are accounts people kept and polished across all that time, because they mattered to them. Engineers play a particular game with any brief: take it seriously for an afternoon and work out what it would cost to build. That game is the whole of what follows.
The design brief
An engineer reading those poems the way they read a client brief would jot down four requirements. Two of them are one question each. The other two turn out to depend on which mode the craft is in when you ask, and that is the whole shape of the thing.
Lift without wings
Hovering is expensive. A small two-seat helicopter holds itself up on roughly 100 kilowatts, about forty kitchen kettles boiling at once, and it pays that bill by throwing a gale of air at the ground. The brief asks for the same lift with nothing visibly thrown. Whatever the craft would push against, the kettle bill would still come due, so the first question is energy density: how much go is packed into each kilogram carried on board.
Silence
Aircraft noise is mostly the sound of air being shoved. The rule usually quoted here is that every push needs a push-back, and it is worth knowing where that rule comes from, because it is a result rather than a starting assumption. Emmy Noether showed in 1918 that conservation laws are consequences of symmetries: momentum is conserved because empty space looks the same from one place to the next. That symmetry holds beautifully in the flat space a helicopter flies through, so in that setting the rule is as firm as rules get.
It is a statement about a setting, though, not about all settings. In general relativity, where spacetime itself is curved and moving, there is no single global momentum ledger to balance; the bookkeeping is local, and physicists have argued about how to write it down since the 1910s. Miguel Alcubierre published a metric in 1994 in which a craft is carried by a reshaped region of space rather than by anything thrown out the back. Nothing is expelled in that picture, so there is nothing to hear. Whether such a region can be made, and what it would cost, is exactly the open work.
Meanwhile quiet thrust in flat space is already built. An ion engine, which throws charged atoms instead of air, flies on deep-space probes today and runs close to silent; its push is about the weight of a couple of sheets of paper resting on your palm. Strong quiet thrust is the gap in this layer. Silence without any exhaust at all is a different question, and it is a question about which layer you are working in.
Inertial handling, which is not the same question as momentum
Start with what a g-force is, because the usual account has it the wrong way round. The feeling is not made by going fast, and it is not made by changing speed either. It is made by a surface pushing on part of you and that push travelling through the rest: the seat shoves your back, your back shoves your ribs, your ribs shove your organs. The squeeze running through you is what you feel, and it is what does the damage.
Gravity works the other way. It pulls on every particle of you at the same rate at the same instant, so nothing inside you gets pressed against anything else and no squeeze builds. In general relativity an object following a geodesic, the straightest path available through curved spacetime, has zero proper acceleration by definition: an accelerometer bolted to it reads zero however hard the coordinates say it is turning. Einstein got there in 1907 and called it the happiest thought of his life. For someone in free fall there is, in their immediate neighbourhood, no gravitational field to feel.
The space station sits about 408 kilometres up, where Earth still pulls at 8.67 metres per second squared, 0.88 of a g. It goes round the planet at 7.67 kilometres a second and closes the circle every 93 minutes. The crew feel about a millionth of a g. That is a ratio near 900,000 to one between the acceleration they are under and the load they carry, and nothing aboard is doing the damping. The shape of the path is. What little they do feel comes from air drag, machinery and each other, which is where the word microgravity comes from.
Alcubierre's 1994 metric says exactly this about its own occupants, and says it in one line. The craft moves on a geodesic, so in his words the proper acceleration along the spaceship's path will always be zero
, however the coordinate acceleration runs. The zero is not a device bolted into the cabin. It falls out of being carried instead of shoved, which is the same reason the crew upstairs are floating. Anything carried by the shape of space inherits that for free.
Now the size of the problem for a craft that is shoved. Along the head-to-foot axis, the one that drains blood out of the head, unprotected tolerance sits around 4 to 6 g with a measured population mean near 4.7 and a spread from about 2.7 to 7.8. Colour goes first, then vision narrows to a tunnel, then it goes altogether, then consciousness follows; at 6 g and above with no straining that whole sequence can run in four to six seconds. An anti-g suit, which inflates bladders around the legs and belly, buys about one more g. The straining manoeuvre, forced exhalation against a shut throat every three seconds with legs, abdomen and back clamped hard, buys three or four more. Suit and straining together reach about nine, and that is active, exhausting work rather than protection you can sit inside.
| What | Load | Axis and duration | What it is |
|---|---|---|---|
| An astronaut in orbit | 0.88 g of coordinate acceleration, about 0.000001 g felt | continuous, for months | Carried, not shoved. The felt load is drag and machinery, not the orbit. |
| An airliner's structural limit | +2.5 g, −1.0 g | any axis, manoeuvring | 14 CFR 25.337 sets the positive limit between 2.5 and 3.8 by weight; the structure is proved to 1.5 times that under 25.303, so 3.75 g. |
| A person, no equipment | about 4.7 g, spread 2.7 to 7.8 | head to foot, sustained | Greyout, tunnel vision, blackout, then loss of consciousness. Average incapacitation runs about 12 seconds afterwards. |
| A pilot with suit and straining | about 9 g | head to foot, sustained | The F-16 was drawn to +9.0 g and −3.0 g, matched to the person rather than to what the airframe could take. |
| An unlimited aerobatic aircraft | ±10 g solo, ±8 g with two aboard | either direction, seconds | The Extra 300's structural rating. Above the fighter range the binding part is the occupant. |
| John Stapp, 10 December 1954 | 46.2 g peak, about 20 g averaged across the stop | chest to back, 1.4 seconds | 632 mph in five seconds on nine rockets at Holloman, then a water brake. Almost every capillary in both eyes burst; his sight came back over the following day and never fully returned. |
| Kenny Bräck, 12 October 2003 | 214 g | mixed, milliseconds | Texas Motor Speedway at about 220 mph, recorded by the car's own crash recorder. Survivable because the structure stretched the stop out over milliseconds. |
Eighty-odd years of work, from Wilbur Franks's anti-g suit of 1941 onward, has moved sustained human tolerance from roughly five g to roughly nine. That is a factor under two, bought with a pressure garment and a breathing technique, and nothing in it touches inertia. The 214 g end of the ladder is a different trick again: a carbon survival cell, a head restraint and a soft barrier, all of them working by stretching the stopping distance so the same change in speed arrives over a longer time. Those are the best inertial dampeners anyone has built, and every one of them manages the surface push rather than removing it.
What free fall does not cancel is the part worth writing into a specification. The equivalence principle is local: over a real body the field is not perfectly even, and the difference between the path your head is on and the path your feet are on shows up as a stretch. Two metres of person at the surface of a neutron star of 1.4 solar masses and 12 kilometres across gets about 4.3 × 108 metres per second squared of difference between the two ends, roughly 44 million g of pull, against 214 g as the worst spike anyone has taken and come back from. The same two metres at the horizon of Sagittarius A*, the 4.3 million solar mass hole at the centre of the galaxy, gets about 0.0011 metres per second squared, roughly a ten-thousandth of a g, less than the tug of a lift starting. Tidal pull at a horizon falls off as the square of the mass, so bigger is gentler.
One more thing belongs on the record before anyone prices such a machine. Inertia is measured to extraordinary precision and its behaviour is described exactly. What is not settled is why matter resists being accelerated at all. Ernst Mach argued in 1883 that a body's inertia comes from its relation to all the other matter around it, with the distant stars supplying the reference. Einstein named that Mach's principle in 1918, in Annalen der Physik, and listed it beside general covariance and the equivalence principle; he hoped general relativity would deliver it. Gödel's 1949 rotating universe and the Brans-Dicke theory of 1961 are two of the reasons the question is still argued in the current literature. Gravity Probe B, launched 2004 and reported in 2011, measured rotating matter dragging inertial frames around the Earth, which is a Machian flavour of effect and shows local mass shaping local inertial structure. It does not show where inertia comes from.
There is a second surprise underneath. The Higgs field is usually named as the thing that gives matter its mass, and for the electron that is right. For a proton it accounts for a sliver. A 2018 lattice calculation by Yi-Bo Yang and colleagues, published in Physical Review Letters, splits the proton's mass four ways: quark energy 33 per cent, glue field energy 37 per cent, the trace anomaly 23 per cent, and the Higgs-derived quark mass term 9 per cent, and that last figure includes the strange quark. Nearly all the inertia of ordinary matter is binding energy in the strong interaction. Anyone talking about trimming a craft's mass in flight is mostly talking about that binding energy, which is a different job from the one the word mass usually suggests.
Two things are known to apply force to every particle of a body at once: gravity, and inertia itself. A magnetic field acting on a diamagnetic body comes close, which is why the frog further down floated instead of being crushed. That is the seam, and the bench works it.
Heat, which is a bill for one mode and not for the others
Heat is not a property of going fast. It is the bill for shoving air out of the way with the front of the hull, and it comes due when the hull is the thing doing the shoving.
The arithmetic is short. Air brought to rest against a surface turns its motion into heat, and the temperature rise is the speed squared divided by twice the heat capacity of air, about 2,010 in metric units. Add that rise to the surrounding air, which up in the stratosphere sits near minus 55 degrees Celsius. At Mach 2 the sum lands at about 120 degrees. Concorde's nose ran to 127 degrees at Mach 2.02 cruise, and 127 was the number the crew had to slow down for. At Mach 5 the same sum returns about 1,050 degrees. The SABRE engine's precooler was built for incoming air at around 1,000 degrees at exactly that speed, chilling it to minus 150 in a hundredth of a second. Two separate checks, both close, on one line of arithmetic.
Push it to orbital speed, 7.8 kilometres a second, and it returns about 30,000 degrees. Nothing on a returning capsule runs that hot, and the reason is worth having: above roughly two kilometres a second the air stops behaving like a gas being squeezed and starts coming apart, so most of that energy goes into pulling molecules to pieces and into light radiated away rather than into the wall. A Shuttle silica tile ran to about 1,260 degrees and the reinforced carbon-carbon at the wing edges to about 1,510. The sum gives the heat carried in the flow; the wall sees a fraction of it. At heats like this the choice of scale hardly matters: kelvin, the unit that counts up from absolute zero rather than from the freezing point of water, always reads 273 higher than Celsius, so at thirty thousand the two agree to within one per cent and one figure covers both.
Now the part that makes this a mode question rather than a hull requirement. Every figure above assumes the hull's leading face is what the air meets. Take that assumption away and the requirement goes with it.
- With a field ahead of the hull, the flow is turned before it arrives. Above about two kilometres a second a good share of that flow is charged, which is what a magnetic or electrostatic barrier acts on; the neutral remainder still arrives, and no barrier of that kind has flown.
- Carried by the metric, the surrounding matter is not being shoved at all. Alcubierre's own paper puts the harsh conditions at the bubble wall, where the radial coordinate approaches the bubble radius, and describes the region around the craft as quiet when the wall is thick and the bubble is larger than the ship. The limit moves from the skin to the wall.
- Out of phase with the surrounding matter, the exchange that heats a surface is the thing that has been changed. What would settle it is a coupling measurement: how strongly the craft and the medium trade momentum and energy, in newtons and watts, at a stated overlap. That is an instrument question rather than a materials one.
So an ablator or a tile is what a craft carries for the mode where the hull does the work, and it rides along as mass in the modes where it does not. The bench below counts exactly that.
| Shielding | What it handles | The figures |
|---|---|---|
| Whipple bumper | Debris and micrometeoroids | Fred Whipple, 1947. A thin outer sheet, a gap, then the real wall. The sheet shocks the projectile into a spreading cloud so the wall meets it across an area instead of at a point. Closing speeds in orbit run 10 to 20 km/s. |
| Stuffed Whipple, Nextel and Kevlar | Debris, more of it per kilogram | A ceramic and aramid blanket in the gap. The station's version stops a 13 mm aluminium sphere at 7 km/s: about 3.1 grams carrying 76 kilojoules, the energy of a family car at 36 km/h landing on a spot a centimetre across. The station carries over 100 different shield fits. |
| Ablative shield, PICA | Entry heat, once | Carbon felt filled with phenolic resin at 0.27 g/cm3. It chars and sheds, carrying the heat away with the material. Stardust came back on 15 January 2006 at about 12.8 km/s, the fastest return flown, with peak stagnation heating near 1,100 W/cm2. |
| Reusable ceramic tiles | Entry heat, repeatedly | LI-900 silica, about 94 per cent air by volume, good to about 1,260 °C, with carbon-carbon to about 1,510 at the wing edges. The Shuttle carried roughly 24,300 tiles and inspected them between flights. Starship carries about 18,000 rated to about 1,400 °C, 10.5 tonnes of them on the first block. |
| Transpiration cooling | Sustained heat, adjustable | Coolant driven out through a porous wall, so a gas film sits between the flow and the metal. Oxford's Mach 7 flat-plate runs got continuous film coverage 5 to 20 mm downstream at blowing ratios from 0.04 to 0.3 per cent, with helium far more effective than air. Used in rocket chambers since the 1960s, not yet as an entry shield. |
| Plasma barrier, a mini-magnetosphere | Charged particles | Ruth Bamford's group at Rutherford Appleton showed the working part is a charge-separation sheath at the cavity edge, not brute magnetic bending, and photographed the cavity in a plasma tunnel. Their worked craft: a 3 m superconducting coil, 8,000 turns at 700 A, peak field about 6.4 T, 16 kW all up, 1.5 tonnes. Nothing of the kind has flown. |
| Electrostatic gossamer nodes | Charged particles | Ram Tripathi's 2011 NASA study: deflect rather than stop, using charged aluminised membranes. Design point of 20 m nodes at minus 17 kV against 10 m nodes at plus 35 kV, with the charge also inflating the membranes. Holding potential against the dangerous high-energy particles costs under a milliwatt; the ordinary low-energy solar wind is what runs the power bill up. |
| Plasma window | A pressure difference across an opening | Ady Hershcovitch's arc at Brookhaven holds vacuum against a room with no solid in the way: 2.5 atmospheres across a 3 mm hole, or a four-inch pipe held at 0.032 Torr against open air, about 24,000 to 1 across four centimetres. A 175 keV electron beam passes straight through it. It ran 2,000 hours without maintenance. The aperture is millimetres. |
Carrying more than one mode is ordinary engineering
Before the bench, the design sense behind it. When a vehicle has to work across regimes whose best answers contradict each other, engineers stop looking for one optimum and carry two. Air is free reaction mass low down and absent high up, so you breathe, then you carry. A nozzle shaped for sea level is the wrong shape in vacuum, so you fly two nozzles. Thrust that gets you off a pad is the wrong thrust for a decade of cruise, so you take both. Combustion needs oxygen and a submerged boat has none, so you run diesel, then battery. Every one of these is certificated, flown or sailed hardware. The question is not whether more than one mode is allowed; it is where the changeover sits and what the changeover machinery weighs.
One engine, two cycles
SABRE breathes air below the changeover and runs on carried oxygen above it, through the same combustion chambers. Specific impulse, the measure of how much push you get per kilogram of propellant spent, is about 3,600 seconds breathing and about 460 seconds as a vacuum rocket: an eightfold difference from one engine. The change happens at Mach 5.14 and 28.5 kilometres, where the inlet cone shuts.
The precooler is the hard part and the enabling part: 16,800 thin-walled tubes running a helium loop that drops incoming air from about 1,000 degrees to minus 150 in a hundredth of a second. Frost blocking the tubes was a real failure mode, answered with a methanol de-icer patented in 2015. Reaction Engines went into administration on 31 October 2024. In July 2025 ESA and Frazer-Nash started INVICTUS, taking on former Reaction Engines staff, to fly a reusable Mach 5 precooled vehicle, with a first flight targeted for early 2031.
An engine that changes what it is, in flight
The Pratt & Whitney J58 in the SR-71 is a conventional afterburning turbojet for take-off. Above about Mach 2 to 2.2 six bypass tubes open and dump air from the fourth compressor stage straight into the afterburner, and the propulsive cycle becomes a turbo-ramjet. At Mach 3.2 cruise the engine core itself supplies about 20 per cent of the thrust. Nobody in 1966 called that exotic; it was the answer to an aircraft that had to leave a runway and then cruise at Mach 3.2. It flew until 1998.
The airframe reshaping itself
A supersonic inlet has to slow the air to subsonic before the compressor, and the shock structure that does it only sits right at one speed unless the geometry moves. So it moves. The SR-71's inlet spike translates 26 inches between Mach 1.6 and its Mach 3.2 design point, and at cruise the inlet contributes 54 per cent of the thrust against the engine's 17 and the ejector's 29. Concorde's hinged intake ramps take Mach 2 down to about Mach 0.5 at the compressor face, and the intake assemblies make 63 per cent of the powerplant's net thrust when the shocks sit where they should.
Same chemistry, two geometries
Falcon 9 flies essentially the same Merlin engine on both stages, with a far larger nozzle expansion on the upper one, 165 to 1, because it never runs in atmosphere. Sea level: 854 kilonewtons at 282 seconds. Vacuum: 981 kilonewtons at 348 seconds. Nothing changed but the shape of the bell, and the specific impulse moved 23 per cent. Starship does the same by flying sea-level Raptors and vacuum Raptors side by side on one stage.
Diesel, then battery, then a third mode
The diesel-electric submarine has been a two-mode vehicle since before the First World War, for the plain reason that combustion needs oxygen and the boat has none down there. Surfaced or at snorkel depth the diesels charge the bank; submerged, motors run off stored charge, and the mast and exhaust plume make the changeover a tactical event rather than a convenience. Air-independent propulsion adds a third mode sized for endurance instead of speed: a Gotland-class boat on Stirling AIP stays down at 5 knots for roughly two weeks against a few days on batteries alone.
Chemical departure, ion cruise
Dawn had to enter orbit at Vesta, leave, and enter orbit at Ceres: about 11 kilometres a second of velocity change after launch. Its NSTAR thrusters gave 92 millinewtons at 2.6 kilowatts, throttling from 3,200 seconds of specific impulse down to 1,900 as sunlight thinned, and down to 19 millinewtons at half a kilowatt; twelve 0.9 newton hydrazine engines did the work that had to happen at a particular moment. Hayabusa's ion set delivered 1,400 metres a second on 22 kilograms of xenon across 25,800 thruster-hours. BepiColombo's T6 thrusters throw xenon out above 50 kilometres a second. Neither engine type is the better one; they are different engines for different regimes.
What one mode is bounded by
Tsiolkovsky's equation says the velocity change a rocket gets is its exhaust speed multiplied by the natural logarithm of the mass ratio: what the vehicle weighs full, divided by what it weighs empty. The exhaust speed is set by chemistry and nozzle shape and does not move much; the best flown chemical exhaust runs about 3.73 kilometres a second. Everything else has to come out of that logarithm, and a logarithm is a slow way to buy anything: you multiply the mass ratio to add velocity in even steps.
Low Earth orbit costs about 9.4 kilometres a second of velocity change: 7.8 for the orbit itself, roughly 1.0 to 1.5 lost climbing against gravity, 0.1 to 0.15 to drag, then manoeuvring and margin. Run the equation backwards at 3.73 and the mass ratio comes to about 12.5, which puts 92 per cent of the liftoff mass into propellant. Structure, engines, tanks, shielding, landing gear and payload share the other 8 per cent. That is why rockets stage, and why taking the oxidiser out of the air for part of the climb is worth the trouble of building a precooler.
Starship is the current worked example, and the published figures let anyone check the sums. Super Heavy carries 3,400 tonnes of propellant at 3,675 tonnes gross: a propellant fraction of 92.5 per cent. The upper stage carries 1,500 tonnes at 1,585 gross, 94.6 per cent. The full stack before payload is about 5,260 tonnes, of which 4,900 is propellant, 93.2 per cent. Liftoff thrust is 89.5 meganewtons, enough to hold up about 9,100 tonnes, roughly 4,300 loaded utes. Raptor runs 327 seconds of specific impulse at sea level and 380 in the vacuum version, and both fly on the same vehicle. Staging happens about 141 seconds in, near 64 kilometres. These are the same proportions the Saturn V ran on: 93.8 per cent propellant in the first stage, 92.3 in the second, 89.0 in the third.
What one mode buys is a single propellant, a single tank set, a single engine family, and a cadence that gets cheaper the more often it runs. What it costs shows up in the same arithmetic. At 94.6 per cent propellant, every tonne of dry mass on the upper stage comes straight out of payload, and the heat shield is dry mass: about 18,000 tiles, 10.5 tonnes on the first block. Run the equation on that stage at 380 seconds. Empty, the mass ratio is 18.6 and the ideal velocity change is 10.9 kilometres a second. With 100 tonnes aboard the ratio drops to 9.1 and the figure to 8.2, against 9.4 to reach orbit; the booster supplies about 3.4 of it. The margins are thin by design and the vehicle is flying.
The universal shelf
Here is the quietly wonderful part, and it is measured, not imagined. Starlight carries a barcode. Split any star's light with a prism and thin dark lines appear at exact colours; each element prints its own line pattern, as personal as a fingerprint. In 1868 astronomers found a pattern in sunlight that matched nothing known on the ground and named the element helium, after the Greek word for the Sun. It took another 27 years for anyone to find helium here on Earth. The Sun's shopping list was read from 150 million kilometres away, a trip a passenger jet would need about twenty years of flying to make, and it was correct.
Every star measured so far prints its barcode from the same set of elements. So the shelf is universal: an engineer near any star would be shopping from the same stock list of 118 kinds of atom, and the entire list fits on one classroom wall chart. No secret ingredients have turned up anywhere light has been read.
Which means any real edge would come from arrangement, not ingredients. Three arrangements already show what that edge looks like.
Perfect crystals
A metal's weakness mostly lives in its flaws. Tiny flawless whiskers of iron hold near 13 gigapascals of pull before they part, where a gigapascal is a measure of how hard something is pulled or pressed, close to ten thousand times the pressure of the air around you. Ordinary soft iron from the same shelf gives way near 0.2, about sixty times less. Same atoms, tidier stacking, very different material.
Chosen isotopes
An isotope is the same element at a slightly different weight. Diamond grown from carbon of a single weight sheds heat noticeably better than natural diamond. The chemistry is identical; the sorting does the work.
Metamaterial lattices
A metamaterial is a material whose trick is its shape, not its chemistry. Built as a fine lattice, an ordinary substance can steer sound or light in ways no plain block of the same stuff can. See how that works.
The back of the shelf
If the brief forced a shopping list, an engineer might write down four lines. None of them is a stock item you could order this afternoon; the measurements around each one are real.
Metallic hydrogen
Squeeze hydrogen hard enough and the prediction, first made in 1935, is that it would turn into a shiny metal. Labs press samples between two shaped diamonds, point to point. Ranga Dias and Isaac Silvera at Harvard reported a reflective sample at 495 gigapascals in Science in 2017, a squeeze beyond the pressure at the centre of the Earth, like parking two thousand utes on a space the size of a fingernail. The sample was lost when a diamond shattered during a follow-up measurement, other groups have questioned how the pressure was worked out, and nobody has repeated it. If it could be kept stable at everyday pressure, predictions say it would carry current without loss and hold about twenty times the energy of the fuel a Space Shuttle main engine burned.
Room-temperature superconductor
A superconductor carries electric current without losing any of it as heat. The warmest confirmed one at everyday pressure works near minus 140 degrees (133 kelvin), colder than the coldest outdoor air ever recorded on Earth. Under crushing pressure, hydrogen-rich compounds are reported superconducting near minus 20 degrees (253 kelvin), close to a kitchen freezer. A superconductor happy in a warm room at everyday pressure would let a craft's wiring and magnets run without waste; none is confirmed so far.
Orbit-grown single crystals
Labs aboard the space station grow crystals in free fall, and many come back more even than the same recipe grown on the ground; reports describe fewer flaws. Today's best space-grown pieces sit closer to a sugar grain than a brick. A hull grown as one flawless piece would need that process scaled up millions of times, which nobody has attempted.
Lattices thinner than soap film
Graphene is a sheet of carbon one atom thick; a stack of three hundred thousand sheets would be about as thick as a page of a book, and each sheet is far thinner than the wall of a soap bubble. Tiny flawless pieces measure stronger for their weight than any steel. A hull skin assembled layer by chosen layer would be a material with no catalogue entry yet.
The design bench
A shopping list is cheap. A sum is not. So here is the brief turned into a bench with a parts catalogue: eight hulls, eight ways of carrying power, eight ways of pushing against the world, and eight ways of shielding. Every option names something built or measured, with its figures attached.
The craft has one hull and one set of dimensions. Everything else is chosen mode by mode, because the brief describes three jobs and they do not share an answer.
- In air. The craft holds itself up against gravity and shoves atmosphere aside. Reaction mass is free because it is all around; heat is real because the hull is what does the shoving.
- Off the air. Nothing to push against and nothing to heat the skin. Anything thrown had to be carried up first, and debris arrives at ten to twenty kilometres a second.
- Carried. The craft is moved by the geometry around it rather than by throwing anything. No reaction mass, no felt load on the crew beyond the gradient across the cabin, and the flow is not arriving at the hull.
Each mode picks its own power source, its own way of pushing and its own shielding, and each is checked against its own requirement. Then the craft has to carry all of it at once, including the parts that earn their keep in one regime and ride along in the others. That last line is where most of the interest sits.
The whole craft, carrying all three
- Hull mass: the skin's mass per square metre, times that skin area.
- Shielding mass: each different shielding choice is built once and counted once, even where two modes share it. Area shields scale with the skin; the plasma barrier is a fixed 1.5 tonnes and a 16 kilowatt draw.
- What each mode has to move: in air, the whole craft against gravity at 9.81 metres per second squared. Off the air, the whole craft at whatever acceleration the course-change slider is set to. Carried, the geometry moves it and there is no thrust rating to compute.
- Power for that mode: force divided by the way of pushing's newtons per kilowatt, plus any draw the shielding wants.
- Power plant mass: that power divided by the source's watts per kilogram. Two modes on the same source share one plant, sized for the larger of the two demands.
- The closing test: every plant has to be lifted along with everything else, in every mode, so the sum runs round again until it settles. It settles only if the plants can hold up the craft and themselves. Where it does not settle, the bench sizes the plants for the bare craft and says by what factor they overrun.
- What the crew feel: a surface push passes through the body in full, so the felt load is the whole acceleration. A body force, applied to every particle at once, leaves only the variation across the cabin, which is the evenness slider.
Keep wandering
Anomalous Metamaterials
Build alien or ancient craft skins from crystal geometry, phase memory and field-responsive interfaces.
Metamaterials
Where shape beats chemistry: lattices that steer sound, light and load.
Zero point
The Casimir force is measured. What a machine would have to do with it is the open part.
Sci-fi lab
Named fictional technology, given a spec sheet and a price in real units.
Frontiers
The paths from today's records towards materials nobody has held yet.